Quality Assurance of Data Transmission in Queuing Networks

Abstract

With the development of high-speed communication networks, the so-called property of self-similarity of flows has an increasing impact on the quality of service. From a practical point of view, this can be explained by the high variability of traffic intensity and, as a consequence, the high receipt of packets to the network node at a high data rate, which leads, due to the limitation of the buffer, to packet losses. For a long time, it was believed that the traffic of the local network is described by the classical Poisson distribution. Telephone networks were originally built on the principle of channel switching, and computer networks are usually based on the principle of packet switching, but the calculation methods have remained virtually the same. Packets at high speed of their movement on a network arrive on a node not separately, and the whole pack. Traffic in such networks has ripples, which increases the likelihood of congestion in the network nodes, which lead to buffer overflows and cause losses and / or delays. Pulsations lead to differences in the speed of information flows, in which the ratio of the maximum value to the minimum speed is tens of times. At the same time, it turned out that in multiservice networks, the number of events in a given time interval depends on previous, very distant events. This means that at large scales of a multiservice network, traffic has the property of self-similarity, i.e. it looks qualitatively the same at any sufficiently large scales of the time axis.